Adaptive neuro-fuzzy classifier for ‘Petit Mal’ epilepsy detection using Mean Teager Energy

An epileptic seizure is an abnormal harmonious neural activity in the brain characterized by the presence of spikes in the electroencephalographic patterns. Petit Mal is a common form of epilepsy (a neurological disorder resulting in recurrent seizures) in children. An automated detection of Petit Mal seizures assists the neurologists in effective diagnosis, thereby enabling proper on-time treatment of epileptic patients. The seizures were mainly detected previously using time-frequency analysis and artificial neural networks. The proposed approach utilizes the abnormality found in the EEG of a Petit Mal patient to create an efficient detection system involving five-level wavelet decomposition based features and adaptive neuro-fuzzy interference system as the classifier. Mean Teager Energy is the only feature used in the proposed method. Unlike previous approaches, the proposed work does not suffer from large noise and sensitivity, thus giving an accuracy of 100% and run-time delay of less than 30 seconds for 100 epochs. This is a tremendous improvement over other methods.